RESUMEN
INTRODUCTION: The Consortium to Establish a Registry for Alzheimer's Disease (CERAD) is a renowned cognitive test battery, which has been extended in its German version to the CERAD-Plus including tests of executive functions and processing speed. The most commonly used total score (TS) is based on the restricted CERAD version and reflects the sum of selected raw-values (Chandler et al., 2005). The CERAD-Plus extensions might be of particular diagnostic utility for cognitive assessments in Parkinson's Disease (PD), as executive functions and processing speed belong to the most vulnerable domains in PD. OBJECTIVE: The aim was to develop a CERAD-TS based on the extended CERAD-Plus' age-, gender-, and education-corrected z-scores and to evaluate its diagnostic accuracy compared to the established CERAD-Chandler-TS. METHODS: Baseline data of n = 679 patients with PD (69% male, n = 277 PD without cognitive impairment, n = 307 PD-MCI, n = 95 PD-D) from the multicenter, prospective DEMPARK/LANDSCAPE study were analyzed. ROC-analyses were conducted for four different TS that were either based on the original CERAD or CERAD-Plus, on raw-values or z-scores, and equally-weighted or based on factor scores. AUC-comparisons were conducted to determine the best yet most parsimonious TS. RESULTS: The newly designed CERAD-Plus-TS based on equally-weighted z-scores outperformed both the CERAD-Chandler-TS and cognitive screening instruments when differentiating between individuals with PD of varying cognitive impairment (0.78 ≤ AUC ≤ 0.98). CONCLUSION: Results suggest a high relevance of non-amnestic subscales for the cognitive assessment in PD populations. The proposed CERAD-Plus-TS needs further validation. The extensions might offer diagnostic potential for non-PD populations as well.